5:08 p.m.
Dear all,
I want to call an R function for drawing stacked histograms using
gretl's foreign block. As only numeric values (scalar, matrix) can be
passed to R, I first write a temporary txt-file containing the string
for the path+file-name for the plot into $dotdir. Then I call the R
function and read-in the txt-file.
Using R, this function works fine. But it doesn't work properly when
calling from inside gretl. I obtain the error:
<error>
External command failed
*** error in function Rhist, line 8
</error>
and in the CLI error
<CLI>
gretl_R_function_exec: R_catch failed on HIST_R
</CLI>
Here is the script:
<hansl>
set verbose off
set R_functions on
function void Rhist (matrix X, string fname)
# MAIN FUNCTION
string sdir = sprintf("%s/tmpRhist.txt", $dotdir)
outfile @sdir --write --quiet # write the path+file name into a
txt-file
printf "%s\n", fname
outfile --close
ret = R.HIST_R(X) # Should "R.HIST_R(X)" also just work?
end function
foreign language=R
HIST_R <- function(X) {
library(Cairo)
library(cairoDevice)
# FIXME
# for (i in 1:dim(X)[2]){
# #v <- as.numeric()
# H$i <- numeric( hist(X[,i]) ) #,breaks=seq(mn,mx,bin))
# }
H1 <- hist(X[,1])
H2 <- hist(X[,2])
H3 <- hist(X[,3])
# read-in "path+file name" string
fname <- paste(readLines("/home/at/.gretl/tmpRhist.txt"),
collapse=" ")
CairoPNG(filename = fname, width = 600, height = 600, pointsize = 26,
bg = NA, res = NA)
par(family="Times", las=1)
plot(H1, col=rgb(0,0,0,1/2), xlab="", main="")
plot(H2, add=T, col=rgb(0.5,0.5,0.5,1/2), xlab=NULL, ylab=NULL
,main=NULL)
plot(H3, add=T, col=rgb(1,1,1,1/2), xlab=NULL, ylab=NULL ,main=NULL)
dev.off()
ret<-1
return(ret)
}
end foreign
#---------
# EXAMPLE
#---------
matrix M = mnormal(200,3)
string fname = sprintf("%s/Rhistcheck.png", $dotdir)
Rhist(M, fname)
</hansl>
Best,
Artur
5:40 p.m.
Am 08.03.2018 um 18:08 schrieb Artur T.:
Dear all,
I want to call an R function for drawing stacked histograms using
gretl's foreign block. As only numeric values (scalar, matrix) can be
passed to R, I first write a temporary txt-file containing the string
for the path+file-name for the plot into $dotdir. Then I call the R
function and read-in the txt-file.
fname <-
paste(readLines("/home/at/.gretl/tmpRhist.txt"),
Don't know if this would solve your problem (actually it probably
shouldn't), but gretl's dotdir path is available inside the foreign R
environment as 'gretl.dotdir', as mentioned (only in a footnote) in ch.
39 of the guide.
Also, maybe it's safer to define the function HIST_R before you use it,
that is, move the foreign block up? (But I'm not sure if it's actually
necessary here.)
good luck,
sven
11:24 p.m.
On Thu, 8 Mar 2018, Artur T. wrote:
I want to call an R function for drawing stacked histograms using
gretl's foreign block. As only numeric values (scalar, matrix) can
be passed to R, I first write a temporary txt-file containing the
string for the path+file-name for the plot into $dotdir.
Ah, we've assumed up till now that anyone calling an R function
within gretl will just be interested in numerical stuff. However,
it's not difficult to pass gretl strings to R (he says, trusting
that R, as gretl, does everything in UTF-8). So that's now enabled
in git.
Then I call the R function and read-in the txt-file.
Using R, this function works fine. But it doesn't work properly when
calling from inside gretl. I obtain the error:
<error>
External command failed
*** error in function Rhist, line 8
</error>
Not sure what precisely was going wrong there, but anyway, it's no
longer necessary to create an intermediate string-passing function.
I'll append a working example below, but I should state one thing
first (I'll post a more comprehensive note on this to gretl-users).
With current gretl, please DO NOT write any material to "dotdir"
that you wish to survive after the current gretl session has ended.
dotdir is intended for temporary files, and we now clean it up on
exit. If you want to specify a directory for writing files that
should persist, use $workdir rather than $dotdir.
That said, here's a revision of your script that works with current
git:
<hansl>
set verbose off
set R_functions on
foreign language=R
HIST_R <- function(X, fname) {
library(Cairo)
library(cairoDevice)
H1 <- hist(X[,1])
H2 <- hist(X[,2])
H3 <- hist(X[,3])
CairoPNG(filename=fname, width=600, height=600, pointsize=26, bg=NA, res=NA)
par(family="Times", las=1)
plot(H1, col=rgb(0,0,0,1/2), xlab="", main="")
plot(H2, add=T, col=rgb(0.5,0.5,0.5,1/2), xlab=NULL, ylab=NULL, main=NULL)
plot(H3, add=T, col=rgb(1,1,1,1/2), xlab=NULL, ylab=NULL, main=NULL)
dev.off()
ret <- 1
return(ret)
}
end foreign
# example call
matrix M = mnormal(200, 3)
fname = sprintf("%s/Rhistcheck.png", $workdir)
rret = R.HIST_R(M, fname)
</hansl>
The above writes Rhistcheck.png OK but there's one oddity: there's a
momentary flash of what I presume is an R plot window. Do you know
how to prevent that?
Allin
4:46 p.m.
Thanks Sven and Allin for your help!
Am 09.03.2018 um 00:24 schrieb Allin Cottrell:
On Thu, 8 Mar 2018, Artur T. wrote:
> I want to call an R function for drawing stacked histograms using
> gretl's foreign block. As only numeric values (scalar, matrix) can be
> passed to R, I first write a temporary txt-file containing the string
> for the path+file-name for the plot into $dotdir.
Ah, we've assumed up till now that anyone calling an R function within
gretl will just be interested in numerical stuff. However, it's not
difficult to pass gretl strings to R (he says, trusting that R, as
gretl, does everything in UTF-8). So that's now enabled in git.
> Then I call the R function and read-in the txt-file.
>
> Using R, this function works fine. But it doesn't work properly when
> calling from inside gretl. I obtain the error:
>
> <error>
> External command failed
> *** error in function Rhist, line 8
> </error>
Not sure what precisely was going wrong there, but anyway, it's no
longer necessary to create an intermediate string-passing function.
Cool, that's good to know.
That said, here's a revision of your script that works with
current git:
<hansl>
set verbose off
set R_functions on
The above writes Rhistcheck.png OK but there's one oddity: there's a
momentary flash of what I presume is an R plot window. Do you know how
to prevent that?
Thank you for this updated script, Allin. It works fine. However, here
(ubuntu 17.10.) the figure appears and actually stays permanently. I
can't even close it without closing gretl. For the moment this doesn't
bother me. And no, I did not find an option to avoid the figure showing
up... :-/
Best,
Artur
1:42 a.m.
On Fri, 9 Mar 2018, Artur Tarassow wrote:
Thank you for this updated script, Allin. It works fine. However,
here
(ubuntu 17.10.) the figure appears and actually stays permanently. I
can't even close it without closing gretl. For the moment this doesn't
bother me. And no, I did not find an option to avoid the figure showing
up... :-/
This is maybe off-topic, but I took a further look. I know that R is
capable of producing very nice data graphics, but the language to
control graphics seems to me a rabbit-hole of doom. Comments beow
relate to R-3.4.3 on Fedora 26.
The two packages you loaded in your script at
http://lists.wfu.edu/pipermail/gretl-users/2018-March/013035.html ,
namely Cairo and cairoDevice, are not complementary, they're
competitors (not unusual for R packages). So I've tried each of them
individually (script below). Both of them work when the script is
run in gretlcli -- and in that context cairoDevice is better in one
way: it respects your request to use the Times font, which is
ignored by the Cairo package. However, when run via the gretl GUI
the cairoDevice variant fails. Neither of the packages respects
your par() setting of "las=1" (axis labels always horizontal).
The graphics system seems to depend on a bunch of inscrutable global
variables over which the packages have an insecure degree of
control. Maybe it's unfair, but I have a sense of package writers
not quite knowing what they're doing in this area. Or maybe the API
has been changed since the packages were written.
As for the flashing (or persistent) X11 window showing a simplified
version of the plot we're trying to write to PNG, it seems to be a
feature of R that you can't turn that off. However, by sticking
"dev.off()" before the call to Cairo() I was able (on my system at
any rate) to get the X11 window just to flash, and not persist, when
using gretl_x11. (Using gretlcli it always just flashed.)
<hansl>
set verbose off
set R_functions on
foreign language=R
HIST_R1 <- function(X, fname) {
# package Cairo 1.5-9 (2015-09-26)
library(Cairo)
H1 <- hist(X[,1])
H2 <- hist(X[,2])
H3 <- hist(X[,3])
dev.off()
# dimensions in pixels
CairoPNG(filename=fname, width=600, height=600,
pointsize=14)
par(family="Times", las=1)
plot(H1, col=rgb(0,0,0,1/2), xlab="", main="")
plot(H2, add=T, col=rgb(0.5,0.5,0.5,1/2))
plot(H3, add=T, col=rgb(1,1,1,1/2))
dev.off()
0
}
HIST_R2 <- function(X, fname) {
# package cairoDevice 2.24 (2017-01-06)
library(cairoDevice)
H1 <- hist(X[,1])
H2 <- hist(X[,2])
H3 <- hist(X[,3])
dev.off()
# dimensions nominally in inches
Cairo_png(filename=fname, width=7, height=7,
pointsize=14)
par(family="Times", las=1)
plot(H1, col=rgb(0,0,0,1/2), xlab="", main="")
plot(H2, add=T, col=rgb(0.5,0.5,0.5,1/2))
plot(H3, add=T, col=rgb(1,1,1,1/2))
dev.off()
0
}
end foreign
# example calls
matrix M = mnormal(200, 3)
# does not respect choice of font family via par()
printf "Calling R.HIST_R1\n"
fname = sprintf("%s/Rhist1.png", $workdir)
scalar rret = R.HIST_R1(M, fname)
print rret
# fails when called via gretl GUI, but OK in gretlcli
printf "Calling R.HIST_R2\n"
fname = sprintf("%s/Rhist2.png", $workdir)
scalar rret = R.HIST_R2(M, fname)
print rret
</hansl>
Allin
9:23 a.m.
Am 13.03.2018 um 02:42 schrieb Allin Cottrell:
On Fri, 9 Mar 2018, Artur Tarassow wrote:
> Thank you for this updated script, Allin. It works fine. However, here
> (ubuntu 17.10.) the figure appears and actually stays permanently. I
> can't even close it without closing gretl. For the moment this doesn't
> bother me. And no, I did not find an option to avoid the figure
> showing up... :-/
The two packages you loaded in your script at
http://lists.wfu.edu/pipermail/gretl-users/2018-March/013035.html ,
namely Cairo and cairoDevice, are not complementary, they're competitors
(not unusual for R packages). So I've tried each of them individually
(script below). Both of them work when the script is run in gretlcli --
and in that context cairoDevice is better in one way: it respects your
request to use the Times font, which is ignored by the Cairo package.
However, when run via the gretl GUI the cairoDevice variant fails.
Neither of the packages respects
your par() setting of "las=1" (axis labels always horizontal).
The graphics system seems to depend on a bunch of inscrutable global
variables over which the packages have an insecure degree of control.
Maybe it's unfair, but I have a sense of package writers not quite
knowing what they're doing in this area.
I've started to work a bit using R, and my impression is indeed that
there are packages out there which are just a black box. Also the
manuals are either not up to date or superficial.
As for the flashing (or persistent) X11 window showing a simplified
version of the plot we're trying to write to PNG, it seems to be a
feature of R that you can't turn that off. However, by sticking
"dev.off()" before the call to Cairo() I was able (on my system at any
rate) to get the X11 window just to flash, and not persist, when using
gretl_x11. (Using gretlcli it always just flashed.)
Thank you, Allin. The dev.off() before the call to Cairo() or also
plot() does the job! I think that using R for some advanced graphs by
means of the foreign-block is a nice feature. Still, gnuplot is also
quite powerful.
Best,
Artur
5:26 p.m.
On Tue, 13 Mar 2018, Artur Tarassow wrote:
Am 13.03.2018 um 02:42 schrieb Allin Cottrell:
> The two packages you loaded in your script at
> http://lists.wfu.edu/pipermail/gretl-users/2018-March/013035.html , namely
> Cairo and cairoDevice, are not complementary, they're competitors (not
> unusual for R packages). So I've tried each of them individually (script
> below). Both of them work when the script is run in gretlcli -- and in that
> context cairoDevice is better in one way: it respects your request to use
> the Times font, which is ignored by the Cairo package. [...]
Footnote after a little poking around. It seems that both of these
packages are out of date and redundant. This variant of your plot
function works just as well or better, using the built-in png()
function:
foreign language=R
HIST_R <- function(X, fname) {
H1 <- hist(X[,1])
H2 <- hist(X[,2])
H3 <- hist(X[,3])
dev.off()
png(filename=fname, width=600, height=600,
pointsize=24, type="cairo-png")
par(family="Times", las=1)
plot(H1, col=rgb(0,0,0,1/2), xlab="", main="")
plot(H2, add=T, col=rgb(0.5,0.5,0.5,1/2))
plot(H3, add=T, col=rgb(1,1,1,1/2))
dev.off()
0
}
end foreign
Allin
12:04 p.m.
Am 13.03.2018 um 18:26 schrieb Allin Cottrell:
On Tue, 13 Mar 2018, Artur Tarassow wrote:
> Am 13.03.2018 um 02:42 schrieb Allin Cottrell:
>
>> The two packages you loaded in your script at
>> http://lists.wfu.edu/pipermail/gretl-users/2018-March/013035.html ,
>> namely Cairo and cairoDevice, are not complementary, they're
>> competitors (not unusual for R packages). So I've tried each of them
>> individually (script below). Both of them work when the script is
>> run in gretlcli -- and in that context cairoDevice is better in one
>> way: it respects your request to use the Times font, which is
>> ignored by the Cairo package. [...]
Footnote after a little poking around. It seems that both of these
packages are out of date and redundant. This variant of your plot
function works just as well or better, using the built-in png() function:
Ah, that's good to know that there is no need to load additional
external packages!
Thank you.
Artur
2785
days inactive
2791
days old
7 comments
4 participants
participants (4)
-
Allin Cottrell -
Artur T. -
Artur Tarassow
-
Sven Schreiber